Double Porosity Finite Element Method For Borehole Modeling

Rock Mech. Rock Engng.
(2005) 38 (3), 217–242

DOI 10.1007/s00603-005-0052-9
Double Porosity Finite Element Method

for Borehole Modeling
By
J. Zhang1 and J.-C. Roegiers2
1
CIRES, The University of Colorado, Boulder, CO, U.S.A.
2
Mewbourne School of Petroleum and Geological Engineering,
The University of Oklahoma, Norman, OK, U.S.A.
Received June 21, 2004; accepted January 24, 2005
Published online March 15, 2005 # Springer-Verlag 2005
Summary
This paper considers the mechanical and hydraulic response around an arbitrary oriented borehole
drilled in a naturally fractured formation. The formation is treated as a double porosity medium
consisting of the primary rock matrix as well as the fractured systems, which are each distinctly
different in porosity and permeability. The poro-mechanical formulations that couple matrix and
fracture deformations as well as fluid flow aspects are presented. A double porosity and double
permeability finite element solution for any directional borehole drilled in the fractured porous
medium is given. Compared with conventional single-porosity analyses, the proposed double-
porosity solution has a larger pore pressure in the matrix and a smaller tensile stress in the
near-wellbore region. The effects of time, fracture, mud weight, and borehole inclination in the
double-porosity solution are parametrically studied to develop a better understanding of physical
characteristics governing borehole problems.
Keywords: Double porosity, finite element method, inclined borehole, poro-mechanics, fractured
porous media.
1. Introduction
Field observations have revealed a need for a better and more comprehensive method
to model borehole stability, since the exploration and production of hydrocarbons now
occur in ever more difficult geological settings (Maury and Zurdo, 1996; Willson
et al., 1999); such as in naturally fractured media, in shaley formations, and at great
depths. In fractured porous formations, borehole instability has been of major concern
due to potential rock movements along fractures at the borehole wall. In the case of
shaley formations, not only does the state of stress play a role, but also the properties
and interactions between the shale and the drilling fluid. At great depths, the rock can
218 J. Zhang and J.-C. Roegiers
become more ductile and behave as an elastoplastic material. In all those cases, the
design approach via conventional elasticity is not suitable.
Borehole modeling using linear elasticity was first used to predict the stability of a
vertical borehole subjected to a non-hydrostatic far-field stress and constant borehole
fluid pressure (Hubbert and Willis, 1957). Then, Goodman (1966) analyzed stress
distributions around circular openings and described weak plane effects using a
numerical method. The elastoplastic numerical analysis for underground openings
was given by Kovari (1977). Early analyses for a borehole of arbitrary trajectory in
linear elastic media were presented (Fairhurst, 1968; Bradley, 1979). Later on, a semi-
analytical model that took into account the influence of rock anisotropy on inclined
borehole stability was developed (Aadnøy, 1987). When the formation is saturated, the
effective stress field around a borehole is strongly modified by the pore pressure,
which is an important factor affecting borehole stability. Assuming a vertical borehole
with a plane strain deformation geometry, a poroelastic solution was developed in
the case of a non-hydrostatic stress field (Detournay and Cheng, 1988). For inclined
boreholes drilled in isotropic media a poroelastic analytical solution was derived
earlier by applying the generalized plane strain concept (Cui et al., 1997a). The
stability of boreholes in saturated rocks was also studied by using the finite element
method (Aoki et al., 1993; Charlez, 1999). A pseudo three-dimensional finite element
program for coupling anisotropic, nonlinear poroelasticity was formulated to simulate
inclined wellbore problem in porous media (Cui et al., 1997b).
However, most petroleum reservoirs are situated in fractured porous formations
(Pruess and Tsang, 1990). In fact, in order to drain the reservoir, inclined and hor-
izontal wells must be drilled in such fractured porous formations – the pay zones.
Therefore, traditional borehole modeling methods cannot completely satisfy this
requirement; and a new scheme needs to be developed to model boreholes drilled
in the fractured porous medium. This study considers such a medium as the combined
effects of solid rock deformation and fluid flow in both the rock matrix and fractures
by using a double porosity and double permeability geomechanical approach.
2. Double Porosity Poro-Mechanical Model

In the double porosity approach, the naturally fractured reservoir is classified as a
system containing two different physical domains. The primary rock matrix contains
a large volume of fluid but has a rather low permeability; while the fractures constitute a
small volume but have the ability to transmit a large portion of the total flow through
the reservoir. In early approaches (Barenblatt et al., 1960; Warren and Root, 1963;
Kazemi, 1969), the reservoir was considered as two overlapping continua: matrix and
fractures. Flow between the matrix and the fractures was accounted for by the introduc-
tion of source functions. However, the effects of stresses and deformations on both the
matrix and fractures were not considered. Then, a coupled flow-deformation approach
within double porosity poroelastic media was presented (Aifantis, 1997; Wilson and
Aifantis, 1982), and a constitutive model to define response of a fissured medium was
given (Elsworth and Bai, 1992). Later, a series of papers were published to study the
fluid flow and solute transport in multiple porosity media, such as Bai et al., 1993; Bai
and Roegiers, 1997; Bai et al., 1999a; Moutsopoulos et al., 2001 and Alboin et al., 2002.
Double Porosity Finite Element Method 219
Fig. 1. Naturally fractured and ideal double porosity reservoirs
In this study, the naturally fractured reservoir is assumed to be an ideal double

porosity, double permeability model as shown in Fig. 1. In this model the system
including matrix blocks and fractures is considered as a continuous medium.
The relationship between total stress (ij) and effective stress (0ij ) was given by
Terzaghi (1943) and Biot (1941). For double porosity media, the effective stresses can
be expressed as:
maij ¼ 0maij þ ma pma ij ;
ð1Þ
frij ¼ 0frij þ fr pfr ij ;
where subscripts ‘ma’ and ‘fr’ represent matrix and fractures, respectively; p is the
pressure; is the effective stress coefficient; and, ij is the Kronecker delta.
For separate but overlapping porous media, the linear elastic constitutive relation-
ships among the effective stresses and strains are defined as (Elsworth and Bai, 1992;
Bai et al., 1999b):
"makl ¼ Cmaijkl 0maij ;
ð2Þ
"frkl ¼ Cfrijkl 0frij ;
where Cmaijkl and Cfrijkl are the compliance tensors for the rock matrix (subscript ma)
and fracture systems (subscript fr), respectively. The detailed expressions for the
tensors are listed in Appendix I.
The total strain due to the elastic deformation in each of the systems is given by:
"kl ¼ "makl þ "frkl ð3Þ
Substituting the double effective law, i.e., Eq. (1) into Eq. (2) while combining
with Eq. (3) and noting maij ¼ frij ¼ ij , one obtains:
ij ¼ Dmfijkl ð"kl þ Cmaijkl ma pma ij þ Cfrijkl fr pfr ij Þ: ð4Þ
The combined elasticity matrix Dmfijkl is defined explicitly in a three-dimensional
geometry for an isotropic medium as:
Dmfijkl ¼ ðCmaijkl þ Cfrijkl Þ1 ; ð5Þ
where the detailed expression for Dmfijkl is given in Appendix I.
In general, the stress-strain behavior of rocks is non-linear, especially in the case
where external loads exceed the elastic strength of material and the material becomes
plastic. The elastoplastic stress-strain can be described by:

s ¼ Dep
mfijkl «; ð6Þ
where Dep p p
mfijkl ¼ Dmfijkl Dmfijkl , Dmfijkl is the plastic modulus tensor.
The total strain increment is assumed to be the sum of the elastic and plastic
components; i.e.
« ¼ «e þ «p ð7Þ
where «, «e and «p are the total, elastic and plastic strain increments, respec-
tively. For more details about perfect plasticity solution, refer to Zhang (2002).
For the separate but overlapping model, the double effective stress law (Eq. (1))
needs to be considered, and the combined elasticity matrix Dmfijkl as well as other
related elastic constants need to be introduced. Then, the governing equations for solid
deformation and fluid phase in the dual-porosity poromechanical formation can be
written as (Zhang, 2002):
Gmf ui;jj þ ðmf þ Gmf Þuk;ki þ ma Dmfijkl Cmaijkl pma;i þ fr Dmfijkl Cfrijkl pfr;i ¼ 0;
kma @"kk @pma
pma;kk ¼ ma Dmfijkl Cmaijkl ma þ !ðpfr pma Þ þ qma ;
@t @t ð8Þ
kfr @"kk @pfr
pfr;kk ¼ fr Dmfijkl Cfrijkl fr !ðpfr pma Þ þ qfr ;
@t @t
where is the relative compressibility representing the lumped deformability of the
fluid and the solid; u is the solid displacement; "kk is the total body strain; ! is the
transfer coefficient (Warren and Root, 1963); s is the fracture spacing; qma and qfr
are the applied fluid flux; mf is Lame’s constant for the combined double porosity
medium; and, mf ¼ Dmfijkl =ð1 þ Þð1 2Þ.
3. Finite Element Discretization of the Poroelastic Solution

The first step in solving the coupled problem of fluid flow and solid deformations is to
discretize the problem domain by replacing it with a collection of nodes and elements
referred to as the finite element mesh. The values of the material properties are usually
assumed to be constant within each element but are allowed to vary from one element
to the next; making it possible to simulate nonhomogeneous problems. The second
step in the finite element method is to derive an integral formulation for the governing
equations. This leads to a system of algebraic equations that can be solved for values
of the field variable at each node in each mesh. The method of weighted residuals is
used for the fluid flow and solid deformation modeling. The Galerkin method is then
used whereby the weighting function for a node is identical to the shape function used
to define the approximate solution.
3.1 Shape Function

Interpolation or shape functions are used to map the element displacements and fluid
pressures at the nodal points.
For the fluid pressure approximation at phase i, one has:

p ¼ Np i i ð9Þ
where N is the shape function for the fluid pressure and solid displacement.
At the nodal level, for four-point two-dimensional elements:
X
4
pi ¼ Nj pij ð10Þ
j¼1
For the eight-point three-dimensional elements:

X
8
pi ¼ Nj pij ð11Þ
j¼1
For eight-point three-dimensional elements, the shape function vector for pressure can
be given in following forms:
1
N1 ¼ ð1 Þð1 Þð1
Þ
8
1
N2 ¼ ð1 þ Þð1 Þð1
Þ
8
1
N3 ¼ ð1 þ Þð1 Þð1 þ
Þ
8
1
N4 ¼ ð1 Þð1 Þð1 þ
Þ
8 ð12Þ
1
N5 ¼ ð1 Þð1 þ Þð1
Þ
8
1
N6 ¼ ð1 þ Þð1 þ Þð1
Þ
8
1
N7 ¼ ð1 þ Þð1 þ Þð1 þ
Þ
8
1
N8 ¼ ð1 Þð1 þ Þð1 þ
Þ;
8
where , and
represent local coordinates; and 1 1, 1 1 and
1
1.
A similar expression for the approximation in mapping nodal displacements can be
described as:
u ¼ Nu: ð13Þ
At the nodal level, for four-point two-dimensional elements:
X
4 X
4
ux ¼ Nj uxj ; uy ¼ Nj uyj : ð14Þ
j¼1 j¼1
For eight-point three-dimensional elements:

X
8 X
8 X
8
ux ¼ Nj uxj ; uy ¼ Nj uyj ; uz ¼ Nj uzj ; ð15Þ
j¼1 j¼1 j¼1
where u is the nodal displacement vector. For simplicity, the superscript ‘’ indicates
the finite element approximations that are omitted in the subsequent description.
Strains within a single element are related to nodal displacements through the
derivatives of the shape functions as:
« ¼ Bu; ð16Þ
where B is the strain-displacement matrix.
The generalized plane strain solutions maintain compatibility with the primary
unknown terms equivalent to the three-dimensional formulation, but geometrically they
are not related to the z-coordinate, similar to the two-dimensional cases. With reference to
the finite element formulation, the major differences among the generalized plane strain,
the plane strain and the three-dimensional situation are exhibited in the strain-displace-
ment matrix B. In a three-dimensional geometry the matrix B can be expressed as:
2@ 3
@x 0 0
60 @ 07
6 @y 7
6 @ 7
6 0 0 @z 7
B¼6 6 @ @ 0 7N;
7 ð17Þ
6 @y @x 7
60 @ @ 7
4 @z @y 5
@ @
@z 0 @x
It is well-known in plane strain problems, i.e., in an x-y plane, that the displace-
ment and the shear stresses are restricted along the z-direction. In the generalized
plane strain scenarios, however, these restrictions are removed. As a result, the number
of tensor components for stresses and strains are identical to that of a three-dimen-
sional setting. In a general generalized plane strain formulation, it is assumed that
boundary conditions in the form of surface tractions, pore pressure, displacements,
and normal flux, do not change along the z-direction. As a result, the displacements,
stresses, strains and pore and fracture pressures are only functions of x, y and time t
(Bai et al., 1999a). For the generalized plane strain formulation, B can be written as:
2@ 3
@x 0 0
60 @ 07
6 @y 7
60 0 07
6 7
B ¼ 6 @ @ 0 7N ð18Þ
6 @y @x 7
6 @ 57
4 0 0 @y
@
0 0 @x
3.2 Conservation Equations

The general force equilibrium equation, in terms of nodal variables for a double
porosity medium in generalized plane strain domain, is given by (Zhang, 2002):
ð ð ð ð
BT Dmf « d
þ ma BT Dmf Cma mNpma d
þ fr BT Dmf Cfr mNpfr d
¼ Nf dS

S
ð19Þ
where
represents the integral domain; C is the compliance matrix defined in
Appendix I with Cma ¼ Cmaijkl, Cfr ¼ Cfrijkl; Dmf is the matrix expression of Dmfijkl; f
is the vector of applied boundary tractions; S is the domain surface on which the sur-
face traction f is applied.
Substituting Eq. (16) and dividing through by t, the momentum balance equation
(Eq. (19)) in the finite element form can be expressed as:
ð ð ð
@u @p @p
BT Dmf B d
þ ma BT Dmf Cma mN ma d
þ fr BT Dmf Cfr mN fr d

@t
@t
@t
ð
@f
¼ N dS ð20Þ
S @t
or,
@u @p @p @F
K þ R1 ma þ R2 fr ¼ ; ð21Þ
@t @t @t @t
where detailed expressions of the coefficients are given in Appendix II.
Using the Gaussian quadrature method (Zienkiewicz, 1977), the double porosity
mass balance equations (last two equations in Eq. (8)) in the finite element forms can
be given for each system.
For the rock matrix system:
ð ð ð
1 @u @p
T T
N kma rN d
pma ¼ ma N mDmf Cma B d
ma NT N d
ma

@t @t
ð
ð
þ ! NT N d
Dp þ NT N d qma ð22Þ

For the fracture system:

ð ð ð
1 @u @p
NT kfr rN d
pfr ¼ fr NT mDmf Cfr B d
fr NT N d
fr

@t @t
ð
ð
! NT N d
Dp þ NT N d qfr ð23Þ

where Dp ¼ pfr pma , mT ¼ (1 1 1 0 0 0); is the domain surface on which the fluid
flux q is applied; and Biot’s effective stress coefficients, , can be evaluated as:
Ksk
ma ¼ 1
Ks
ð24Þ
Ksk
fr ¼ 1 ;
Kfr
where K and K are the bulk moduli of the skeleton for the matrix blocks and the
sk sk
fractures, respectively; Ks and Kfr are the bulk moduli of the solid grains and fractures,
respectively; and, the relative compressibilities, , can be written as:
nma ma nma
ma ¼ þ
Kf Ks
nfr fr nfr ð25Þ
fr ¼ þ ;
sKn sKn
in which Kf and Kn are the bulk modulus of the fluid and the normal stiffness of the
fractures, respectively; n is the porosity; and, s is the fracture spacing.
Equations (22) and (23) can be written as the finite element forms:
@u @p
M1 þ ðQ L1 Þpma Qpfr N1 ma ¼ Qma ð26Þ
@t @t
@u @pfr
M2 Qpma þ ðQ L2 Þpfr N2 ¼ Qfr ; ð27Þ
@t @t
where detailed expressions of the above coefficients are listed in Appendix II.
Equations (21), (26) and (27) represent a set of differential equations in time and
can be expressed in matrix form as follows:
2 32 3 2 3 2 3 2 dF 3
0 0 0 u K R1 R2 u dt
6 76 7 6 7d6 7 6 7
4 0 Q L1 Q 54 pma 5 þ 4 M1 N1 0 5 4 pma 5 ¼ 4 Qma 5:
dt
0 Q Q L2 pfr M2 0 N2 pfr Qfr
ð28Þ
The discretization in space has been completed; Eq. (28) now represents a set of
differential equations in time.
3.3 Finite Element Discretization in Time

Using a fully implicit finite difference scheme in the time discretization domain, such that:
8 tþDt
>
> du 1 tþDt
>
> ¼ ðu ut Þ
>
> dt t
>
< tþDt
dpma 1 tþDt
¼ ðpma ptma Þ ð29Þ
>
> dt t
>
>
>
> tþDt
: dpfr ¼ 1 ðptþDt pt Þ
>
dt t fr fr
and substituting Eqs. (29) into Eq. (28), the finite element equations in the matrix form
for a double porosity poroelastic medium can be expressed as follows:
2 32 3tþt
K R1 R2 u
1 6 76 7
4 M1 ðQ L1 Þt N1 Qt 54 pma 5
t
M2 Qt ðQ L2 Þt N2 pfr
2 32 3t 2 F 3tþt 2 F 3t
K R1 R2 u t t
1 6 76 7 6 7 6 7
¼ 4 M1 N1 0 54 pma 5 þ4 Qma 5 4 0 5 : ð30Þ
t
M2 0 N2 pfr Qfr 0
There are 5 unknowns (ux, uy, uz, pma, pfr) and 5 equations per node; therefore,
displacements and pressures can be solved. In addition, strains and stresses can be
obtained through the following equation and Eq. (4).
1
"ij ¼ ðui;j þ uj;i Þ: ð31Þ
2
3.4 Stress Conversion for an Inclined Borehole

For an inclined borehole with its axis inclined with respect to the principal axes of the
far-field stresses (see Fig. 2), the following equations can be used to convert the global
Fig. 2. Local and global coordinate systems for an inclined borehole
coordinate (far-field stress coordinate, x0 , y0, z0 ) into the local coordinate (borehole
coordinate, x, y, z) system.
8 9 2 2 3
> Sx > lxx0 l2xy0 l2xz0
>
> >
>
>
> Sy > > 6 l2 0 l2yy0 l2yz0 7 8 9
>
< >
= 6 yx 7 < S x0 =
Sz 6 l2 0 l2zy0 2 7
lzz0 7 S 0
¼6
6
zx ð32Þ
lxz0 lyz0 7
y
>
> S >
xy > lxx0 lyx0 lxy0 lyy0 7: Sz0 ;
>
> > 6
>
> Syz >>
>
4 l 0l 0 l 0l 0 lzz0 lyz0 5
: ; yx zx zy yy
Sxz lzx0 lxx0 lzy0 lxy0 lzz0 lxz0
Fig. 3. Finite element mesh in the local coordinate system

where,
8 9 2 3
< lxx0 lxy0 lxz0 = cos ’x cos ’z sin ’x cos ’x sin ’z
lyx0 lyy0 lyz0 ¼ 4 sin ’x cos ’x 0 5
: 0 ;
lzx l y0 lzz0 cos ’x sin ’z sin ’x sin ’z cos ’z
Sx0 , Sy0, and Sz0 are the far-field stresses; Sx, Sy, Sz, Sxy, Syz and Szx are the local borehole
coordinate stresses; ’x is the angle between the global and local coordinates (Fig. 2);
and, ’z is the borehole deviation. After the conversion, the finite element analysis can
be worked out in the local coordinate system, i.e., the section perpendicular to the
borehole axial direction (Fig. 3).
4. Model Validation
The formulations presented in the foregoing section are coded using the Fortran
language in the pseudo-three-dimensional and time domain using four-node rectan-
gular elements. Some analytical problems, such as elastic and poroelastic analytical
solutions for inclined borehole problems, have been examined to validate the com-
puter program.
The geometric loading for an inclined borehole problem is depicted in Fig. 2. The
Cartesian coordinate system (x0 y0 z0 ) is chosen to coincide with the principal axes of the
in-situ compressive stresses, respectively, designated as Sx, Sy and Sz. The initial
formation pore pressure is denoted by p0. The local coordinate system (Fig. 2) is
formed by a rotation of the azimuth angle, ’x , about the x0 -axis, and then by an
inclination of the zenith angle, ’z , from the z0 -axis toward the z-axis.
Then, at the local coordinate system the boundary conditions at the far-field
(r ! 1) are characterized by the normal stresses:
0x ¼ Sx ; 0y ¼ Sy ; 0z ¼ Sz
and the shear stresses:
0
xy ¼ Sxy ; y0 ¼ Syz ; xz0 ¼ Sxz ;
as well as the matrix and fracture pore pressures:
p0ma ¼ p0ma ; p0fr ¼ p0fr ;
where the superscript ‘0’ indicates the virgin state.
In the following analysis, the in-situ stresses and initial pore pressures were (Cui et al.,
1997a): Sx0 ¼ 29 MPa, Sy0 ¼ 20 MPa, Sz0 ¼ 25 MPa, and p0 ¼ p0ma ¼ p0fr ¼ 10 MPa. The
wellbore inclination is defined as ’x ¼ 0 and ’z ¼ 70 . The wellbore radius was
R ¼ 0.1 m. The load at the wellbore is assumed as being applied instantaneously. In the
local coordinate system (after 70 inclination), these values become (Zhang and Roegiers,
2002; Zhang et al., 2003): Sx ¼ 25.5 MPa, Sy ¼ 20 MPa, Sz ¼ 28.5 MPa, Sxz ¼ 1.3 MPa,
Sxy ¼ Syz ¼ 0 MPa, and p0 ¼ p0ma ¼ p0fr ¼ 10 MPa. The formation materials were assumed
to be isotropic, characterized by the following properties: Biot modulus, M ¼ 15.8 GPa;
Biot’s effective stress coefficient, ma ¼ 0.771; permeability, kma ¼ 1 107 darcy;
and, fluid dynamic viscosity, ¼ 0:001 Pa s. The analytical solution for this particular
Table 1. Parameters for inclined borehole analysis
Parameter Unit Magnitude
Elastic modulus (E) GPa 20.6

Poisson’s ratio () – 0.189
Fracture stiffness (Kn, Ksh) MPa=m 4.821 105
Fluid bulk modulus (Kf) MPa 419.17
Grain bulk modulus (Ks) GPa 48.21
Matrix porosity (nma) – 0.02
Fracture porosity (nfr) – 0.002
Matrix mobility (kma=) m2=Pa s 1010
Fracture mobility (kfr=) m2=Pa s 109
Fracture spacing (s) m 107
The use of extremely large fracture spacing is only for the validation with
the single porosity method.
Fig. 4. Comparison between the finite element and analytical solutions for pore pressure ( ¼ 5.7 )
generalized plane strain poroelastic problem was provided by Cui et al. (1997a). The
corresponding equivalent parameters for the dual-porosity poroelastic model are listed
in Table 1, in which the selection of an exceptionally large fracture spacing, s, denotes
the approximation of a homogeneous single-porosity medium.
Figure 4 represents the pore pressure variations into the rock formation. The
comparative results between the analytical solution and the numerical dual-porosity
solution (for large s) along the radial direction ¼ 5.7 are shown at two different
times; the numerical results appear to agree well with the analytical solution.
For the same data set, the Terzaghi’s effective radial stresses, defined as the
difference between the total radial stress and the pore pressure, are plotted in Fig. 5
for ¼ 5.7 . Except for a slight difference in the near-wellbore region, the analytical
and finite element solutions match well. The tensile region developed at early time is
due to the non-monotonic pore pressure distribution, consistent with the case reported
by Cui et al. (1997b).
Fig. 5. Comparison between the finite element and analytical solutions for Terzaghi’s effective radial stress
( ¼ 5.7 )
Fig. 6. Comparison between the finite element and analytical solutions for total tangential stress along
different radial sections ( ¼ 5.7 , ¼ 84.4 ) and times (t ¼ 1.3 min, t ¼ 21.6 min)
Figure 6 presents the total tangential stresses for two different radial directions
( ¼ 5.7 and 84.4 ) and two different times (t ¼ 1.3 and 21.6 min); excellent agreement
can be seen between the finite element and analytical solutions for the larger times
(t ¼ 21.6 min). The small differences in the near wellbore region at small times are
induced by initial conditions, time step and boundary effects.
Fig. 7. Comparison between the finite element and analytical solutions for total axial stress along different
radial sections ( ¼ 0 , ¼ 90 ) and times (t ¼ 1.3 min, t ¼ 21.6 min)
Also, good matches (refer to Fig. 7) are obtained for the total axial stresses for two
different radial directions ( ¼ 0 and 90 ) and two different times (t ¼ 1.3 and 21.6 min).
For the impermeable case, the double porosity finite element model with extreme-
ly large fracture spacing and zero pore and fracture pressures presents the approx-
imation to the elastic solution (analytical solution in Bradley, 1979). The parameters
used for the finite element model are the same as in the previous calculation except
that a mud pressure (pw ¼ 10 MPa) is applied along the borehole wall. Along the radial
section ( ¼ 30 ), the numerical and analytical solutions for radial and tangential
stresses are compared in Fig. 8. It can be seen that the proposed finite element
Fig. 8. Comparison between the finite element and analytical solutions for total radial and tangential
stresses along the radial sections ¼ 30 for the impermeable model
solutions return an excellent matches with the analytical solutions except for small
discrepancies near the borehole wall.
5. Parametric Analyses for the Double Porosity Model

In this section, examples are given to demonstrate how certain parameters in a double
porosity medium affect the pore pressure and stress results. The in-situ stresses,
initial pore pressures, geometry and material properties used in the ongoing analyses
are identical to those listed Table 1, except for the fracture spacing s ¼ 1 m. All the
results are presented at a borehole inclination angle of ’z ¼ 70 . Figure 9 shows
the far-field and local stress states; it can be seen that the x-direction ( ¼ 0 ) is
the local maximum stress direction and the y-direction ( ¼ 90 ) is the local
minimum one.
In the following parametric analyses, double-porosity, time, fracture, mud weight
and inclination effects are considered. For each analysis, only one specific parameter
is allowed to be varied.
5.1 Double-Porosity Effects

The pore pressure at t ¼ 100 s and ¼ 0 for the single-porosity model as well as the
one for the dual-porosity model are compared in Fig. 10. The difference is evidenced
by the increase in pore pressure in the matrix for the dual-porosity media due to the
associated large fracture compliance. A similar comparison is made for Terzaghi’s
effective radial stresses (Fig. 11). Although the pore pressure induced by dual-porosity
effect increases, the tensile stress does not increase, but decreases (Fig. 11). This is
due to the fact that the total radial stress has a larger increment than the pore pressure
for the dual-porosity medium (Zhang, 2002). Therefore, the effective compressive
stress, which equals to the difference of the total stress and the pore pressure, increases
Fig. 9. State of stress in the borehole local coordinate after 70 inclination
Fig. 10. Comparison of pore pressure for single- and double-porosity models in the maximum stress
direction at t ¼ 100 s
Fig. 11. Comparison of effective radial stress for single- and double-porosity models in the maximum stress
direction at t ¼ 100 s
and the effective tensile stress decreases, which reduces the potential for borehole
spalling (tensile failure).
5.2 Time-Dependent Effects

In Figs. 12 through 14, the pore pressure distributions around the wellbore are pres-
ented at three different times. In the near field, pore pressure concentrations occur at
smaller times (for t ¼ 10 s in Fig. 12 and t ¼ 100 s in Fig. 13) around the minimum
stress direction ( ¼ 90 ) and the pore pressures decrease as time increases, as shown
in Figs. 12 to 14. The pore pressure in the near field is larger as increases, which is
due to Skempton’s effect (Detournay and Cheng, 1993), because at 0 a larger far-field
compressive normal stress prevails (Fig. 9). Note that non-monotonic pressure dis-
tributions and pressure peaks are found at a small distance inside the wall at small
times, which is attributed to the poroelastic effect. At larger times this effect disap-
pears. At large distances, the pore pressure approaches asymptotically the far-field
value of 10 MPa, and as time increases this poroelastic effect becomes negligible.
Fig. 12. Pore pressure distribution around the wellbore at t ¼ 10 s
The Terzaghi’s effective radial stress distributions for four different times are
plotted in Figs. 15 and 16. The results clearly show a tensile region near the wellbore
at small times. The tensile radial stress magnitude in the near field is much larger in
the minimum stress direction ( ¼ 90 ) than that in the maximum stress direction
( ¼ 0 ). This is due to much higher pore pressure values in the minimum stress
direction (refer to Figs. 12 and 13). It can also be seen that the radial stress has a
Fig. 15. Effective radial stress in the maximum stress direction for different times
Fig. 16. Effective radial stress in the minimum stress direction for different times
very strong time effect; i.e., the tensile stress reduces significantly as time elapses.
There is no tensile stress as time becomes large enough.
5.3 Mud Weight Effects

Rock failure is likely to take place when a borehole is drilled with air or with
insufficient mud pressure to support it before casing is placed. However, too large
mud pressures may induce borehole instability by tensile fracturing leading to unac-
ceptable mud losses. In the present analysis, the borehole true depth is assumed to be
1000 m and four values of mud density are examined, i.e., w ¼ 0.006, w ¼ 0.01,
w ¼ 0.02 and w ¼ 0.025 MN=m3 corresponding to mud pressures of pw ¼ 6,
pw ¼ 10, pw ¼ 20, and pw ¼ 25 MPa for a borehole depth of 1000 m, other parameters
remaining the same, as in previous analyses.
Due to the mud pressure, there are no tensile radial stresses inside the borehole
except for the case of a very small mud pressure value (e.g. pw ¼ 6 MPa), as shown in
Fig. 17; instead, a non-monotonic stress distribution appears for higher mud pressures.
Fig. 17. Effective radial stress along the minimum stress direction at t ¼ 100 s for different mud pressures
Fig. 18. Effective tangential stress along the maximum stress direction at t ¼ 100 s for different mud pressures
Furthermore, the increasing mud pressure causes an increase of the compressive radial
stress; this is due to the fact that the mud weight acts on the wellbore wall as an
additional radial stress component.
Figure 18 shows the response of the effective compressive tangential stresses along
the maximum stress direction ( ¼ 0 ) at t ¼ 100 s. It is obvious that the compressive
tangential stress decreases as the mud pressure increases, which reduces the high
stress concentration around the wellbore. It is noted that the effective tangential stress
becomes tensile along the local maximum stress direction for high mud pressures (e.g.
pw ¼ 25 MPa). This illustrates that the borehole will fail in tension or fracturing when
the tensile tangential stress is larger than the formation tensile strength.
5.4 Fracture Stiffness Effects

Figure 19 presents the pore pressure responses for different fracture stiffnesses along
the minimum stress direction ( ¼ 90 ). It is obvious that the pore pressure increases
as the fracture stiffness becomes increasingly smaller. When the stiffness is large
Fig. 19. Pore pressure around the wellbore in the minimum stress direction at t ¼ 100 s for different fracture
stiffnesses
Fig. 20. Radial stress around the wellbore in the minimum stress direction at t ¼ 100 s for different fracture
stiffnesses
enough, representing nearly no deformation in the fractures, the pore pressure no

longer varies with fracture stiffness.
The comparisons of the effective radial stresses indicate that the compressive stress
increases while the tensile stress reduces as the stiffness decreases, as shown in Fig. 20.
The radial tensile stress is insignificant when the fracture stiffness is very small.
5.5 Borehole Inclination Effects

To assess the effect of borehole inclination, the pore pressures responses at t ¼ 100 s,
¼ 90 for four different inclination angles, 0 , 40 , 70 and 90 are examined in
Fig. 21. Note that a hole inclination angle of ’z ¼ 0 represents a vertical hole. It
shows that the inclination decreases the pore pressure magnitude in this example. This
results from the extra pore pressure generated by Skempton’s effect. At ’z ¼ 0 the
relevant local far-field compressive stresses at ¼ 0 are x ¼ 29 MPa, y ¼ 20 MPa
which are changed to x ¼ 25 MPa, y ¼ 20 MPa at ’z ¼ 90 due to the borehole
inclination (refer to Fig. 22).
Fig. 21. Pore pressure distribution at t ¼ 100 s, ¼ 90 for different borehole plunges, ’z
Fig. 22. State of stress in the local coordinate systems after 0 and 90 inclinations
Fig. 23. Effective radial stress distributions, at t ¼ 100 s, ¼ 0 , for different borehole plunges, ’z
Fig. 24. Effective radial stress distributions, at t ¼ 100 s, ¼ 90 , for different borehole plunges, ’z
Figures 23 and 24 explore the influence of hole inclination on radial stresses at

t ¼ 100 s, ¼ 0 and ¼ 90 . The inclination causes a reduction in the compressive
radial stress and an increase in the tensile radial stress. This trend is more pronounced
at the azimuthal angle ¼ 0 , because the local far-field total stress varies from
x ¼ 29 MPa (effective stress 19 MPa) at a hole inclination ’z ¼ 0 to x ¼ 25 MPa
(effective stress 15 MPa) compared to a hole inclination ’z ¼ 90 ; however, at an angle
of ¼ 90 , the local far-field total stress does not change (y ¼ 20 MPa) from a hole
inclination ’z ¼ 0 to ’z ¼ 90 (refer to Figs. 9 and 22).
Figures 25 and 26 demonstrate the effect of borehole inclination on the effective
tangential and axial stresses at the azimuthal angle ¼ 90 . It is clear that at larger
inclination angles the tangential stress decreases and the axial stress increases
because the inclination from ’z ¼ 0 to ’z ¼ 90 causes the local far-field total
stress to vary from x ¼ 29 MPa, z ¼ 25 MPa to x ¼ 25 MPa, z ¼ 29 MPa (see
Fig. 22).
From the above analyses it is obvious that the effects of hole inclination on well-
bore pressures and stresses depend strongly upon the far-field stress magnitudes.
Fig. 25. Effective tangential stress distributions, at t ¼ 100 s, ¼ 90 , for different plunges, ’z
Fig. 26. Effective axial stress distributions, at t ¼ 100 s, ¼ 90 , for different borehole plunges, ’z
By the same token, for the given boundary and far-field stress conditions, the borehole
inclination can be optimized in order to avoid high stress concentrations.
6. Conclusions
A naturally fractured reservoir was modeled by using a double porosity poro-
mechanics approach, in which rock matrix and fractures have different mechanical
and hydraulic parameters. Using separate but overlapping rock matrix and fracture
models subjected to double effective laws, a formulation was presented in which
deformations and fluid flow in the matrix and fractures are fully coupled. The finite
element numerical method was applied to solve these equations, and a double porosity
borehole solution was given. The solution was validated via inclined borehole prob-
lems with known elastic and poroelastic analytical solutions.
Parametric analyses showed that the pore pressure and effective tensile radial
stress decrease as time increases, and that the tensile radial stress disappears when
time is sufficiently large. In addition, a non-monotonic pressure distribution and

tensile radial stress appear at a small distance inside the borehole wall at small
times; these are attributed to poroelastic effects. For different fracture parameters
the pore pressure increases and the tensile stress decreases as the fracture stiffness
decreases. Neglecting fracture=double-porosity effects, the pore pressure near the
wellbore will be underestimated. However, when the spacing and stiffness are
large enough, the double porosity solution approaches the single-porosity solution.
In addition, sufficient mud weight can avoid the tensile radial stress along the
borehole; however, high mud weight pressure can induce tensile tangential stress
at the wellbore which causes wellbore fracturing. The effects of the hole inclina-
tion on wellbore pressures and stresses depend on the far-field stress, and the
borehole inclination can be optimized for reducing stress concentration around
the wellbore.
Acknowledgement
The constructive suggestions and academic exchanges from Dr. M. Bai are gratefully acknowl-
edged. Authors also sincerely thank anonymous reviewers for their helpful suggestions and
comments in improving the original manuscript.
References
Aadnøy, B. S. (1987): Modeling of the stability of highly inclined boreholes in anisotropic rock
formations. Paper SPE 16526 presented at SPE Offshore European Conf., Aberdeen.
Aifantis, E. C. (1977): Introducing a multi-porous medium. Devel. Mech. 9, 201–211.
Alboin, C., Jaffre, J., Joly, P., Roberts, J. E., Serres, C. (2002): A comparison of methods for
calculating the matrix block source term in a double porosity model for contaminant transport.
Comput. Geosci. 6, 523–543.
Aoki, T., Tan, C. P., Bamford, W. E. (1993): Effects of deformation and strength anisotropy
on borehole failures in saturated shales. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 30,
1031–1034.
Bai, M., Elsworth, D., Roegiers, J.-C. (1993): Multi-porosity=multi-permeability approach to the
simulation of naturally fractured reservoirs. Water Resour. Res. 29, 1621–1633.
Bai, M., Elsworth, D., Roegiers, J.-C. (1997): Triple-porosity analysis of solute transport.
J. Contam. Hydrol. 28, 247–266.
Bai, M., Abousleiman, Y., Cui, L., Zhang, J. (1999a): Dual-porosity porous elastic modeling of
generalized plane strain. Int. J. Rock Mech. Min. Sci. 36, 1087–1094.
Bai, M., Meng, F., Elsworth, D., Abousleiman, Y., Roegiers, J.-C. (1999b): Numerical modeling
of coupled flow and deformation in fractured rock specimens. Int. J. Numer. Anal. Meth.
Geomech. 23, 141–160.
Barenblatt, G. I., Zhelstov, I. P., Kochina, I. N. (1960): Basic concepts in the theory of seepage of
homogeneous liquids in fissured rocks. Prikl. Mat. Mekh. 24, 852–864.
Biot, M. A. (1941): General theory of three-dimensional consolidation. J. Appl. Phys. 12,
155–164.
Bradley, W. B. (1979): Failure of inclined boreholes. Trans. ASME 101, 232–239.
Charlez, P. A. (1999): Rock mechanics, vol. 2, Petroleum application. Editions Technip, Paris.
Cui, L., Cheng, A. H.-D., Abousleiman, Y. (1997a): Poroelastic solution of an inclined borehole.
ASME J. Appl. Mech. 64, 32–38.
Cui, L., Kaliakin, V. N., Abousleiman, Y., Cheng, A. H.-D. (1997b): Finite element formulation
and application of poroelastic generalized plane strain problems. Int. J. Rock Mech. Min. Sci.
34, 953–962.
Detournay, E., Cheng, A. H.-D. (1988): Poroelastic response of a borehole in a non-hydrostatic
stress field. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 25, 171–182.
Detournay, E., Cheng, A. H.-D. (1993): Fundamental of poroelasticity. Ch.5, In: Fairhurst, C.
(ed.) Comprehensive rock engineering, vol. 2. Pergamon Press, Oxford.
Elsworth, D., Bai, M. (1992): Flow-deformation response of dual-porosity media. J. Geotech.
Eng. 118, 107–124.
Fairhurst, C. (1968): Methods of determining in situ rock stress at great depth. TRI-68 Missouri
River Div. Corps of Engineering.
Goodman, R. E. (1966): On the distribution of stresses around circular tunnels in nonhomoge-
neous rocks. In: Proc., 1st Int. Congress ISRM, Lisbon, 1966, 249–255.
Hubbert, M. K., Willis, D. G. (1957): Mechanics of hydraulic fracturing. JPT, Trans. ASME 210,
153–166.
Kazemi, H. (1969): Pressure transient analysis of naturally fractured reservoirs with uniform
fracture distribution. Soc. Pet. Eng. J. 9, 451–462.
Kovari, K. (1977): The elasto-plastic analysis in the design practice of underground openings.
In: Gudehus, G. (ed.) Finite elements in geomechanics. John Wiley & Sons, 377–412.
Maury, V., Zurdo, C. (1996): Drilling-induced lateral shifts along pre-existing fractures: a
common cause of drilling problems. SPE Drilling & Completion, March.
Moutsopoulos, K. N., Konstantinidis, A. A., Meladiotis, I. D., Tzimopoulos, C. D., Aifantis, E. C.
(2001): Hydraulic and contaminant transport in multiple porosity media. Transport Porous
Media 42, 265–292.
Pruess, K., Tsang, Y. W. (1990): On two-phase relative permeability and capillary pressure of
rough-walled rock fractures. Water Resour. Res. 26, 1915–1926.
Warren, J. E., Root, P. J. (1963): The behavior of naturally fractured reservoirs. Soc. Pet. Eng. J.
Trans. ASME 228, 245–255.
Wilson, R. K., Aifantis, E. C. (1982): On the theory of consolidation with double porosity. Int. J.
Eng. Sci. 20, 1009–1035.
Willson, S. M., Last, N. C., Zoback, M. D., Moos, D. (1999): Drilling in South America: a
wellbore stability approach for complex geologic conditions. Paper SPE 53940, Latin
America and Caribbean Petroleum Engineering Conference, Caracas, Venezuela.
Terzaghi, K. (1943): Theory soil mechanics. John Wiley & Sons, New York.
Zhang, J. (2002): Dual-porosity approach to wellbore stability in naturally fractured reservoirs.
Ph.D. Dissertation, Univ. of Oklahoma.
Zhang, J., Bai, M., Roegiers, J.-C. (2003): Dual-porosity poroelastic analyses of wellbore
stability. Int. J. Rock Mech. Min. Sci. 40, 473–483.
Zhang, J., Roegiers, J.-C. (2002): Borehole stability in naturally deformable fractured reservoirs –
a fully coupled approach. Paper SPE 77355, SPE Annual Technical Conference & Exhibition.
Zienkiewicz, O. C. (1977): The finite element method. 3rd edn. McGraw-Hill, New York.
Appendix I: Combined Elasticity Matrix and Compliance Matrices

The combined elasticity matrix Dmfijkl is given as (Bai et al., 1999b):
2 3
d1111 d1122 d1133 0 0 0
6 d2211 d2222 d2233 0 0 0 7
6 7
1 6 6 d 3311 d 3322 d 3333 0 0 0 7 7:
Dmfijkl ¼ ð1:1Þ
jDmfijkl j 6
6 0 0 0 d4444 0 0 7 7
4 0 0 0 0 d55555 0 5
0 0 0 0 0 d6666
The compliance matrices are given as follows:
2 3
1 0 0 0
6 1 0 0 0 7
6 7
1 6 0 0 0 7
Cmaijkl ¼ 6 7: ð1:2Þ
E66 0 0 0 2ð1 þ Þ 0 0 7
7
4 0 0 0 0 2ð1 þ Þ 0 5
0 0 0 0 0 2ð1 þ Þ
2 1
3
Kn s 0 0 0 0 0
6 0 1
0 0 0 0 7
6 Kn s 7
6 7
16 0 0 1
0 0 0 7
Cfrijkl ¼ 6 Kn s 7 ð1:3Þ
E6
6 0 0 0 1
Ksh s 0 0 7
7
6 0 0 0 0 1
0 7
4 Ksh s 5
1
0 0 0 0 0 Ksh s
and,
" #
1 2ð1 þ Þ 3 1 1 3 3 2 1 1 2 2
jDmfijkl j ¼ þ þ 2 þ 2 ð1:4Þ
Ksh s E E Kn s E E Kn s E
3 " 2 #
1 2ð1 þ Þ 1 1 2
d1111 ¼ d2222 ¼ d3333 ¼ þ þ 2 ð1:5Þ
Ksh s E E Kn s E
d1122 ¼ d2211 ¼ d1133 ¼ d3311 ¼ d2233 ¼ d3322

" #
1 2ð1 þ Þ 3 1 1 2 2
¼ þ þ 2 ð1:6Þ
Ksh s E E E Kn s E
2 " 2
1 2ð1 þ Þ 1 1
d4444 ¼ d5555 ¼ d6666 ¼ þ þ
Ksh s E E Kn s

3 2 1 1 2 3
2 þ 3 ð1:7Þ
E E Kn s E
where E is the Young’s modulus, and Kn, Ksh are the fracture normal and shear
stiffnesses, respectively.
242 J. Zhang and J.-C. Roegiers: Double Porosity Finite Element Method
Appendix II: Terms in Finite Element Matrices

ð
K ¼ BT Dmf B d
ð2:1Þ

ð
R1 ¼ ma BT Dmf Cma mN d
ð2:2Þ

ð
R2 ¼ fr BT Dmf Cfr mN d
ð2:3Þ

ð
1
L1 ¼ rNT kma rN d
ð2:4Þ

ð
1
L2 ¼ rNT kfr rN d
ð2:5Þ

ð
N1 ¼ ma NT N d
ð2:6Þ

ð
N2 ¼ fr NT N d
ð2:7Þ

ð
Q ¼ ! NT N d
ð2:8Þ

ð
M1 ¼ ma NT Dmf Cma mB d
ð2:9Þ

ð
M2 ¼ fr NT Dmf Cfr mB d
ð2:10Þ

ð
F¼ Nf dS ð2:11Þ
S
ð
Qma ¼ NT Nqma d ð2:12Þ
ð
Qfr ¼ NT Nqfr d ð2:13Þ
Author’s address: Dr. Jincai Zhang, Cooperative Institute for Research in Environmental
Sciences, The University of Colorado, Boulder, CO 80309-0216, U.S.A.; e-mail: jincai@
cires.colorado.edu

Double Porosity Finite Element Method For Borehole Modeling

Diunggah oleh

Informasi Dokumen

Deskripsi Asli:

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Double Porosity Finite Element Method For Borehole Modeling

Diunggah oleh

Hak Cipta:

Format Tersedia

Rock Mech. Rock Engng.

(2005) 38 (3), 217–242

Double Porosity Finite Element Method

2. Double Porosity Poro-Mechanical Model

Fig. 1. Naturally fractured and ideal double porosity reservoirs

In this study, the naturally fractured reservoir is assumed to be an ideal double

plastic. The elastoplastic stress-strain can be described by:

3. Finite Element Discretization of the Poroelastic Solution

3.1 Shape Function

For the fluid pressure approximation at phase i, one has:

For the eight-point three-dimensional elements:

For eight-point three-dimensional elements:

3.2 Conservation Equations

For the fracture system:

3.3 Finite Element Discretization in Time

3.4 Stress Conversion for an Inclined Borehole

Fig. 2. Local and global coordinate systems for an inclined borehole

Fig. 3. Finite element mesh in the local coordinate system

Table 1. Parameters for inclined borehole analysis

Parameter Unit Magnitude

Elastic modulus (E) GPa 20.6

5. Parametric Analyses for the Double Porosity Model

5.1 Double-Porosity Effects

5.2 Time-Dependent Effects

Fig. 12. Pore pressure distribution around the wellbore at t ¼ 10 s

Fig. 13. Pore pressure distribution around the wellbore at t ¼ 100 s

Fig. 14. Pore pressure distribution around the wellbore at t ¼ 1000 s

5.3 Mud Weight Effects

5.4 Fracture Stiffness Effects

enough, representing nearly no deformation in the fractures, the pore pressure no

5.5 Borehole Inclination Effects

Figures 23 and 24 explore the influence of hole inclination on radial stresses at

time is sufficiently large. In addition, a non-monotonic pressure distribution and

Appendix I: Combined Elasticity Matrix and Compliance Matrices

d1122 ¼ d2211 ¼ d1133 ¼ d3311 ¼ d2233 ¼ d3322

Appendix II: Terms in Finite Element Matrices

Anda mungkin juga menyukai